Peer-to-peer architecture for content-based music retrieval on acoustic data

In traditional peer-to-peer search networks, operations focus on properly labeled files such as music or video, and the actual search is often limited to text tags. The explosive growth of available multimedia documents in recent years calls for more flexible search capabilities, namely search by content. Most content-based search algorithms are computationally intensive, making them inappropriate for a peer-to-peer environment. In this paper, we discuss a contentbased music retrieval algorithm that can be decomposed and parallelized efficiently. We present a peer-to-peer architecture for such a system that makes use of spare resources among subscribers, with protocols that dynamically redistribute load in order to maximize throughput and minimize inconvenience to subscribers. Our framework can be extended beyond the music retrieval domain and adapted to other scenarios where resource pooling is desired, as long as the underlying algorithm satisfies certain conditions. Categories and ...

• Type: Conference
• Year: 2003
• Where: WWW
• Authors: Cheng Yang
• Keywords: Content-based Music Retrieval, Internet Technology, Music Retrieval Algorithm, Peer-to-peer Search Networks, WWW 2003